➡️ Exact delivery dates may vary, so some features may not be immediately available on the publication date. For your specific access dates, please contact your LivePerson account team.
🚨The timing and scope of all changes remain at the sole discretion of LivePerson and are subject to change.
Agent Workspace - New text profanity filter
LivePerson is committed to providing a safe, secure, and professional communication environment for your agents and customers. We are excited to announce the release of our new text profanity filter, an advanced, high-performance solution designed to automatically detect and mask inappropriate or offensive language in real-time, significantly improving conversation quality and agent safety.
This release replaces the existing regular expression filter with a state-of-the-art and performance-optimized text-matching algorithm, giving you unprecedented control over your communication governance.
Key features and benefits
- Advanced profanity masking: Our text-matching algorithm masks profane words in real-time, based on a brand-configured list and masking pattern. This enhances agent safety and comfort by preventing the viewing of abusive or harassing customer messages.
- Self-service configuration: You can manage the filter directly via the Management Console: Enable or disable the feature, Use the default proprietary list of abusive words (~2000 words) or configure your own list (up to 20,000 characters). You can also configure masking patterns and a notification message (up to 200 words). This gives you unprecedented control and flexibility to align moderation logic precisely with your brand’s standards and specific regional requirements.
- Flexible masking behaviors: Configure how messages are masked with three distinct strategies: Word (replaces each word with same-length asterisks), Single (replaces each word with a single asterisk), or Full (replaces the entire message). These masking behaviors ensure your messaging aligns with your internal compliance and communication policies.
- Agent notifications: Agents receive a customizable system message in transcript when profane text is detected and masked in a message, providing immediate awareness, for example, “The message contains profane words and was masked.” This notification provides a warning mechanism allowing agents to manage conversations with abusive language proactively.
➡️ These are UI notifications in the Agent Workspace and not tracked separately.
- Role-based access control: The “View masked data” permission allows agents or agent managers to override the default masking and view the original message content, providing necessary oversight. This supports auditing and quality assurance while maintaining core agent protection.
➡️ For the "View masked data" permission” to function, the Agent Masking feature should be enabled in your account. Please contact LivePerson Support to enable this feature and view this permission.
Agent view
Consumer view
How the filter works
- Filter logic: Messages sent from the agent/customer are checked after they are sent. If profanity is detected, the message is masked automatically for both the agent and the customer.
- Agent notes exclusion: While masking logic is applied to agent messages, it is not applied to agent notes.
- System integration: The message payload will include a new indicator field (e.g., indicators: [ “PROFANITY” ]) to signal that the data contains abusive or profane content as per the configured words at that time. This info is useful to know for reporting purposes; see our August 6, 2025 release note about the new masking indicator reporting in the Messaging Interaction API.
Activation and configuration
- Access the Management Console.
- Search for "Text Profanity Masking."
- Enable the filter, configure your word list, and define the preferred masking pattern.
Important notes: Please allow 5 to 10 minutes for updates to word lists or masking patterns to sync across your system.
- Availability: Supported on all channels and the Agent Workspace.
- Limitations:
- Not currently available on the Agent App.
- Masking logic is not applied to Agent Notes.
- The Profane Words list can contain only lowercase (including numbers) characters, comma-separated up to 20,000 characters.
- The Notification Message is limited to 200 characters.
- UI notifications are not tracked separately in reporting.
- The profanity filter is not supported in Agent Workspace when using Copilot Translate.
Report Center enhancements
Instant secure form confirmations
In this release, we add a feedback loop to the agent workflow in the Report Center. The manager now sees a confirmation message in the Transcript Viewer the moment a secure form is sent or submitted, giving you total confidence that the data arrived safely.
Performance insights in Transcript Viewer
We’ve added four key responsiveness metrics directly to the Transcript Viewer. You can now track speed and efficiency in real time with:
- TTFR & TTFRA: Time to First Response (and Acknowledged)
- ART & ARTA: Average Response Time (and Acknowledged)
Updates to dashboard privacy and sharing for agent managers
We’re making dashboard management more secure with a new permission-based sharing system for agent managers.
What’s changing? If your sharing permissions are turned off:
- The Share button will be deactivated for any dashboard that you own.
- Hovering over the button will show a tooltip explaining your access level and the dashboard's current visibility. See our image below.
- To change a dashboard that you own from public to private, just duplicate it to create a private copy, then delete the old one. It will not be possible to change the state between public-view-only and public-edit.
New hourly precision in reporting
Why settle for daily summaries? We’ve upgraded the date picker with an hour filter, allowing you to drill down into your data with much higher granularity. You can now refine your reports by specific hours, making it easier than ever to spot trends within a single day.
Bug fix: Targeted voice enablement validation
We’ve refined how the application handles voice enablement checks to prevent premature error messages:
- On-demand validation: The account check for voice now triggers only when launching the Transcript Viewer, ensuring users without voice accounts enjoy a clean, error-free application launch.
Copilot Translate enhancements and bug fixes
Extending Copilot Translate to third-party channels
We’ve extended Copilot Translate capabilities to include conversations arriving via third-party channels. This enhancement ensures a consistent, multilingual support experience for agents, regardless of the conversation's origin.
Bug fix regarding secure forms
Previously, in conversations via Web Messaging and third-party channels, in English-language conversations (where no translation is expected), submission or receipt of a secure form incorrectly triggered an attempt at translation. We've fixed this issue.